Bug 33299 – [Gtk] FreeType backend does not respect FC_MATRIX property

• Status: RESOLVED
• Resolution: FIXED
• Component: WebKitGTK

[Dino Morelli]

Created attachment 46018 [details]
sample HTML document illustrating CSS italic issue

Text that should display as italic is not rendering italic. Example HTML document attached.

This is happening in both the Midori and uzbl browsers on Arch Linux 32-bit with libwebkit 1.1.15.4

[Dino Morelli]

Versions of webkit browsers tested on the Arch Linux system:

midori 0.2.2-1
uzbl-git 20091205-1

[Martin Robinson]

The issue here seems to be that we are totally ignoring the FC_MATRIX property of the FcPattern we use to create the font.

[Martin Robinson]

The FC_MATRIX property is *very* lightly documented, but we can observe it in Pango.

[Martin Robinson]

Created attachment 68273 [details]
Patch for this issue

[Alexander Butenko]

ping?

[Martin Robinson]

I'll see if I can find someone to review this today.

[chris fleizach]

Comment on attachment 68273 [details]
Patch for this issue

View in context: https://bugs.webkit.org/attachment.cgi?id=68273&action=review

r=me, but you should address the "new SimpleFontData" thing, even though its existing code

> WebCore/platform/graphics/cairo/SimpleFontDataCairo.cpp:89
>  

shouldn't this be adoptRef or something. Naked news are no longer in favor i believe.

[Martin Robinson]

(In reply to comment #7)

> shouldn't this be adoptRef or something. Naked news are no longer in favor i believe.

Thanks for the review! I don't know if adoptRef is appropriate here because SimpleFontData isn't a descendant of RefCounted. In particular, it seems like 
m_smallCapsFontData is owned only by SimpleFontData.

[Martin Robinson]

(In reply to comment #8)

> Thanks for the review! I don't know if adoptRef is appropriate here because SimpleFontData isn't a descendant of RefCounted. In particular, it seems like  m_smallCapsFontData is owned only by SimpleFontData.

Chris, given this, do you still mind if I land the patch as-is?

[Martin Robinson]

I'm almost 100% certain about the SimpleFontData issue, so I'll land as-is. If this is wrong though, we can roll it out ASAP.

[Martin Robinson]

Committed r69320: <http://trac.webkit.org/changeset/69320>